Escape Per-Seat Fees and Manual Invoice Entry
Invoicing apps charge high per-seat fees for features you rarely use. The real cost is the manual data entry your team performs to work around their limits.
A custom system makes sense when your billing is tied to project milestones, variable rates, or multiple data sources. The complexity comes from mapping your unique business logic into code, not from the technology itself. A services business with consistent project-based billing is a typical 3-week build.
We built a system for a 15-person marketing agency struggling with project-based billing. Their finance admin spent two days a month reconciling timesheets. The system we deployed generates 50+ project invoices in 90 seconds, cutting their monthly close from two days to two hours.
What Problem Does This Solve?
Most small businesses start with QuickBooks or FreshBooks. These are great for simple product sales but fail with service-based billing. Their internal automation cannot handle logic like 'if a task in Asana is marked complete, pull hours from Harvest for that task, apply the project-specific rate, and add it to a draft invoice'. This forces your team into manual CSV exports and VLOOKUPs, which is slow and causes errors.
To bridge this gap, teams turn to integration platforms. The problem is that a single invoice might require 10-15 tasks to pull data, calculate rates, and create a line item. For a firm sending 100 invoices a month, that is 1,500 tasks. This pushes you into a high monthly pricing tier for a single, business-critical workflow that still feels brittle.
The core issue is that these tools are not built for complex, multi-step financial calculations. They can move data from A to B, but they cannot reliably perform the custom math that happens in between. This leads to under-billing, angry clients asking for corrections, and a finance process that depends entirely on one person's spreadsheet skills.
How Does It Work?
First, we connect to your data sources with read-only API access. We use Python with httpx for asynchronous calls to your time tracker, project manager, and CRM. We pull 12 months of historical data to build and test the logic, staging it in a Supabase Postgres database. This initial data sync for a year's worth of records completes in under 5 minutes.
The core billing logic is written as a FastAPI service. This Python code replaces the manual spreadsheet work. It iterates through projects, applies the correct hourly rates per role, checks against project budget caps, and bundles line items. For vague time-tracking entries, we use the Claude API to standardize them into client-friendly descriptions. This single step cut invoice clarification emails by over 50% for one client.
We deploy the FastAPI service on AWS Lambda, so you only pay for compute time when an invoice is actually generated. The service generates each invoice as a PDF in under 300ms. For a client generating 400 invoices per month, the total AWS bill is under $20. There are no per-seat fees or task limits. The entire system runs for a fraction of an off-the-shelf automation tool's monthly cost.
Finally, the system pushes completed draft invoices into your existing accounting software like QuickBooks or Xero via their APIs. We use structlog for detailed logging and configure Slack alerts for any API failures. If the accounting platform is down, the system automatically retries. The entire process to generate and post 100 distinct project invoices takes less than 3 minutes.
What Are the Key Benefits?
Generate 100 Invoices in 3 Minutes
Stop spending days on manual data entry. Run your entire monthly billing cycle in the time it takes to make coffee. No more CSV exports or spreadsheets.
One-Time Build, Under $20/Month to Run
You pay a fixed price for the build and own the system. Monthly AWS Lambda costs are minimal, eliminating per-seat fees that grow with your team.
You Get the Full Source Code in GitHub
We deliver the complete Python codebase to your company's GitHub repository. You are never locked into a platform and can have any developer extend it.
Alerts in Slack Before a Client Notices
Built-in monitoring using structlog sends a detailed alert if an API fails or data is missing. You can fix issues proactively, not after a client complains.
Connects Directly to Asana and Xero
We build direct API integrations to your specific tools. The system pulls data from your project manager and pushes invoices to your accounting platform.
What Does the Process Look Like?
Week 1: Scoping and API Access
You provide read-only API keys for your systems. We deliver a data flow diagram mapping every field from source to final invoice line item for your approval.
Week 2: Core Logic and PDF Generation
We build the Python service that handles your unique billing rules. You receive a sample set of generated PDF invoices with real data for review and feedback.
Week 3: Deployment and Integration
We deploy the system on AWS Lambda and connect it to your accounting software. You receive a technical runbook and we process a live batch of invoices together.
30-Day Monitoring and Handoff
We monitor the system for one month post-launch to handle any edge cases. After this period, you have the full source code, with an optional flat-rate support plan available.
Frequently Asked Questions
- How much does a custom invoicing system cost?
- Pricing depends on the number of data sources and the complexity of your billing rules. A system pulling from two APIs with clear logic is a 2-3 week build. Adding more sources or complex rules, like multi-currency support, can extend the timeline. We provide a fixed-price quote after the discovery call so you know the full cost upfront.
- What happens if an API like QuickBooks is down when we run invoices?
- The system is designed for these failures. Each invoice is processed independently. If the QuickBooks API is unavailable, the job is placed in a retry queue in Supabase. It attempts to post again every 15 minutes. If it still fails after 2 hours, a Slack alert is sent with the invoice details so nothing gets lost or double-counted.
- How is this different from using Stripe Invoicing?
- Stripe Invoicing is excellent for subscriptions or simple per-item billing. It is not designed to pull data from external project management tools and calculate totals based on complex business logic. Our system is the glue. It connects to your tools like Jira or Harvest, performs the calculations, and then can use Stripe's API to create and send the final invoice.
- Can this system create an audit trail?
- Yes. Every generated invoice and its source data are stored in a Supabase table. This creates a permanent, auditable record of exactly why an invoice total is what it is, timestamped and tied directly to the source API payloads. This is often more reliable for audits than tracking changes in a shared spreadsheet.
- How do you handle sensitive API keys?
- We never store credentials in code. All API keys and secrets are managed through AWS Secrets Manager, with access restricted to the Lambda function's execution role. For development, we request read-only, scoped-down permissions so we can only access the endpoints necessary for the invoicing workflow. You can revoke access at any time.
- What if our billing logic or rates change in the future?
- Since you own the code, changes are straightforward. A common update like adding a new service rate is often a single-line change in a configuration file. Any Python developer can make these adjustments. For more significant changes to the billing logic, we can scope a small, fixed-price follow-on project.
Related Solutions
Ready to Automate Your Small Business Operations?
Book a call to discuss how we can implement ai automation for your small business business.
Book a Call